Archive for June 20, 2014

Published BICEP2 paper admits “Unquantifiable Uncertainty”..

Posted in Bad Statistics, The Universe and Stuff with tags , , , , , , on June 20, 2014 by telescoper

Just a quick post to pass on the news that the BICEP2 results that excited so much press coverage earlier this year have now been published in Physical Review Letters. A free PDF version of the piece can be found here.  The published version incorporates a couple of important caveats that have arisen since the original release of the results prior to peer review. In particular, in the abstract (discussing models of the dust foreground emission:

However, these models are not sufficiently constrained by external public data to exclude the possibility of dust emission bright enough to explain the entire excess signal. Cross correlating BICEP2 against 100 GHz maps from the BICEP1 experiment, the excess signal is confirmed with 3σ significance and its spectral index is found to be consistent with that of the CMB, disfavoring dust at 1.7 σ.

Since the primary question-mark over the original result was whether the signal was due to dust or CMB, this corresponds to an admission that the detection is really at very low significance. I’ll set aside my objection to the frequentist language used in this statement!

There is an interesting comment in the footnotes too:

In the preprint version of this paper an additional DDM2 model was included based on information taken from Planck conference talks. We noted the large uncertainties on this and the other dust models presented. In the Planck dust polarization paper [96] which has since appeared the maps have been masked to include only regions “where the systematic uncertainties are small, and where the dust signal dominates total emission.” This mask excludes our field. We have concluded the information used for the DDM2 model has unquantifiable uncertainty. We look forward to performing a cross-correlation analysis against the Planck 353 GHz polarized maps in a future publication.

The emphasis is mine. The phrase made me think of this:


The paper concludes:

More data are clearly required to resolve the situation. We note that cross-correlation of our maps with the Planck 353 GHz maps will be more powerful than use of those maps alone in our field. Additional data are also expected from many other experiments, including Keck Array observations at 100 GHz in the 2014 season.

In other words, what I’ve been saying from the outset.